According to the FBI, about 25% of health insurance claims in the US involve trickery. Fraudulent claims often cite escalating losses and mishaps to obtain payouts. One of the most compelling challenges that insurance companies still face is the detection and prevention of such deceptions. The failure to identify sophisticated insurance frauds results in a loss of USD 40 billion annually. The US Department of Justice is now heavily focused on identifying fraudulent business practices across industries – evident from its 2016 crackdown on health care fraud schemes involving nearly USD 900 million.
Challenges in Detecting Insurance Fraud
Ensuring timely detection of insurance fraud comes with its own set of challenges. The primary concern for both US federal and state departments is that stiff penalties may alienate health service providers. Imposing sanctions has been controversial, as opponents have often raised claims of institutional abuse of power and violations of constitutional rights. Insurance frauds manifest in duplicitous filing of claims to elicit compensation from insurers – using genuine patient information to inappropriately bill for services; falsifying or availing unnecessary services to produce superfluous payment information.
Digital Technology and Changing Dynamics of the Insurance Sector
Developments in digital technology has facilitated the generation of large volumes of usable and analyzable data streams. Yet, poor data management has resulted in inefficient assessment of claims, letting fraudulent cases fall through the cracks. Insurers can gather information from social media accounts of claimants to reveal inconsistencies and misrepresentations while validating claims documentation. But the additional use of data analytics to corroborate circumstantial claims has proven extremely effective.
Social Network Analytics (SNA) is an analytical model used to establish linkages between individuals, locations, and other things, which may or may not have been available to insurers at the time when a claim is made. For instance, let’s assume a car accident involving several individuals, all of whom have shared their addresses and phone numbers with the insurer. It might turn out that one of the addresses already has a claim, or the vehicle is involved in several claims. The SNA tool brings together several analytical methods like organizational business rules, statistical methods, pattern analysis, and network linkage analysis to identify patterns across the data pool. Analyzing these patterns, or clusters, and studying how they link to other clusters helps in detection of fraud. The data repository can contain all kind of public records like foreclosure documents, criminal records, address change memos, and bankruptcy declarations.
Combatting Insurance Fraud with Technology
Insurers have recognized technology to be a valued ally in combatting insurance fraud. Link analysis is one such tool that is used to navigate through extensive sets of insurance data to establish complex relationships. A link analysis based framework can help identify patterns of behavior that indicate the probability of fraud. Predictive modelling, text mining, data visualization, and geographic data mining are some of the other techniques used to check deception.
Insurers are proactively making use of data streams generated by sensors, wearables or personal technology, and other geographic information systems (GIS) to effectively assess risks, and the veracity of claims. For example, Erie Insurance is a US firm that was the first insurance company to seek and receive permission from the Federal Aviation Administration (FAA) to use drones commercially. The company says that the benefits are numerous; from speeding up the claim process and looking at damage without endangering employees, to getting a clearer picture of potentially fraudulent cases.
The Future of Health Insurance
Very recently, several doctors in New York were arrested for being involved in a USD 50 million healthcare fraud suit. The use of analytics platforms have already led to more effective anti-fraud measures and successful apprehension of fraudsters (the 2016 US Department of Justice crackdown mentioned at the outset). The most useful and constructive data streams that health insurers can utilize will primarily come from the deployment of wearable technology in the future. Wearable technology, referred to as ‘fit-tech’ help generate health-related metrics – heart rate and number of steps clocked, among others. Prototypes have already been developed which conduct blood work and ECGs and also automatically administer drug doses. In such a scenario, filing false claims will become increasingly difficult as real-time data will be used by insurers to discern the authenticity of a claim.
What do you think about the potential of analytics with regard to keeping health insurance fraud at bay? Does this technology hold the promise to drive a significant change in the insurance industry?